I am pleased to share LexaChat, a full-stack AI conversational platform built with a dual-mode architecture focused on privacy, scalability, and user control.
LexaChat is designed to go beyond a basic chatbot interface. It combines secure authentication, flexible storage architecture, and modern AI integration into a SaaS-ready system.
Key Features:
Dual-Mode Architecture
• Local Mode – Chats stored securely in browser localStorage with full user control (import, export, delete)
• Session Cloud Mode – JWT-authenticated sessions with MongoDB cloud synchronization
Authentication & Security
• Email verification (OTP-based)
• JWT-secured authentication
• Protected backend routes
• Profile management with avatar upload
• Password reset functionality
Advanced Chat Capabilities
• Prompt editing
• Regenerate response functionality
• Multiple response versions
• Markdown and code rendering
• Copy-to-clipboard actions
• Session search and management
User Experience
• Dark, Light, and Cyber themes
• Web search toggle
• Session Vault dashboard
• Real-time cloud sync indicators
Technical Stack:
Frontend: React + Tailwind CSS
Backend: Node.js + Express
Database: MongoDB
Authentication: JWT
AI Integration: MoE-based model proxy architecture
Deployment: Production-ready hosting
LexaChat reflects my progression from building frontend applications to architecting a complete AI-powered system with authentication, storage abstraction, and scalable backend integration.
The platform is live at:
I welcome feedback from developers, AI enthusiasts, and product builders.
Top comments (0)